[Solved. Probably.] I guess this is a rigging question...?

john_antkowiak

So when you convert a primitive object into a figure, it adds a bone and from there you build a skeleton. I've watched enough YouTube videos to get the basics, but they have nothing to say about what you do when something goes unexpectedly. I can't find anything comprehensive about how to use these tools. Does anyone know what I did wrong to get the bounding box in this screen capture to do whatever it's doing? How do I fix it?

I guess I don't understand a few other things too, like... does the first parent bone have to exert rotational influence on anything, in order for the child bones to do so? Do the child bones have to start where their parent bones end? Can they start at the parent bone's mid-point? Can they be disconnected entirely? Suppose for example I'm building a saddle and I want to rig the stirrups and straps. Nothing else on the saddle needs to move, so do I really need one central bone and a bunch of unused bones reaching down to the top of the stirrup straps? 

Richard Haseltine

I'm not sure on the bounding box.

Bone weight maps are, by default, normalised - the total weights on any vertex for any map (the general Weights, or each of the individual TriAx maps) will add up to 100% - adding weight will remove it from another bone and vice versa. Initially the root node gets all the weight.

john_antkowiak

A post on another subject makes me think @Omniflux might have some idea why this bounding box issue occurs and how I can fix it. I hope tagging you will draw your attention...

john_antkowiak

Or maybe @SickleYield has seen this before? One thing that occurs to me is that the rigging videos I've watched don't talk about the Figure Setup tool so I have never used it. Is there a tutorial on its purpose? As with all the other tools, I'd love something that starts with "This is called the Figure Setup Pane and this is what it's for. And now here's an explanation of every concept, tab, radio button, and check mark available to you."

And BTW @RichardHaseltine thank you for that insight. It took me a while to work out which part of my questions you were addressing but I got it  lol  ))

Omniflux

I have not seen this specific issue with a bounding box before, and have no guess as to why it might have happened (I do know what post you are referring to and why you think I might have an idea).

 

The first thing I would try is selecting the figure and running this in the Script IDE

Scene.getPrimarySelection().invalidateBoundingBoxes(true)

If that doesn't work, I would export the object and then update the base geometry using the exported file (Edit -> Figure -> Geometry -> Update Base Geometry - Update Full Geometry).

john_antkowiak

Thank you for the suggestion, @Omniflux. I had never used Script IDE so that was interesting. But the command seemed to have no effect. Admittedly I was expecting to see an immediate change but I have no idea whether that expectation was valid. Then I updated the base geometry and wound up with two different errors. The first was caused I think because I didn't zero the parameter dials before executing the Update, so it halved the Y and Z scale presets. I started over, zeroed them first, and wound up with the correct shape at least... but the bounding box was even further removed.

Omniflux

Okay, I played around and found one possible explanation for this. If you have any bones that extend beyond the actual boundary of your object, the bounding box will expand to encompass them.

Use the Joint Editor tool and examine each bone of your object.

john_antkowiak

Hmm... That is a possibility but if such a thing is happening, I don't know how to fix it because here's how I build it in the first place:

I wanted a strap of a certain dimension; in this case 12" long by 1" wide by .25" thick. So I created a 12" cube primitive with 12 divisions. If you set the edge of the primitive at X=0,Y=0, Z=0 then the first bone begins at 0,0,0 and ends at 2.54, 0, 0; the next ends at 5.08,0,0; then 7.62,0,0 etc. There's no guesswork in where the next bone ends; the last one should end precisely at the edge of the cube and appears to do so in the planimetric viewport options. Then when the rigging is done I add a second cube of 1" to the scene and a third cube of .25". Then I scale the rigged cube to match each one and wind up with a shape 12" by 1" by .25" that sometimes has a crazy bounding box. But not every time, which confuses me. Sometimes the blue "Assign to Target Group" button on theTool Settings pane for the Geometry Editor tool snaps the bounding box where it belongs, but if it changes anything, it only changes once even though I try it agan with various groups selected. I have no idea what that button is supposed to do or how I'm supposed to be using it - and the bounding box doesn't behave the way I'd expect even after it' been snapped with this button. 

felis

Maybe try to switch to the joint editor tool, so you can see where your bones are.

john_antkowiak

Yikes! you're right, felis. After I updated the geometry and got that completely severed bounding box, I switched to the Joint Editor and found all the bones completely removed from the cube. Previously, the bones have always been where they belong but the bounding box has varied widely, disconnected from the bones. Granted, the bounding box has never been removed from the object in every dimension like it is here in the last screenshot. So this is new. Let me work on that for a bit and see what happens...

john_antkowiak

OK so both your comments helped me figure out what I think I've learned. I think the bones got ripped out because the Export to obj function didn't export the bones but did export the revised dimensions of the cube. That's why updating the geometry zeroed the Scale values of the .duf. That's clever - so now instead of loading a cube shrunk by certain percentages to achieve a 12" x 1" x .25", I now have a strap with those measurements as the zero value. I like it. But the .duf retained the previous values of the joint positions based on the original dimensions of the 12" x 12" x 12" cube. Updating their positions to match the current shape yielded the first screenshot below - note the new bounding box. Then I realized that the height of the bounding box was determined by the orientation of the original cube as well. I'd started by loading a cube primitive with the Y Positive orientation, so when I converted it to a figure, Studio gave it the YXZ property you can see in the Joint Editor tool settings. The tutorial videos for rigging sort of implied that Studio assigned this orientation and it shouldn't be altered; all the child bones had to match it. That's why sometimes I had weird bounding boxes and sometimes I didn't - it depended on the orientation of the original primitive. So even though I had rotated the base bone 90 degrees, the bounding box didn't change. It matched the height of the base bone in its original Y orientation. When I reoriented the top node and all the child bones to ZXY, now I have a bounding box that's a lot closer match.

If you look close, you can see it's still a little bit outside the shape but that's explained by the desired position of the bones. I want the strap to rotate along the edges, not along the center - and the bounding box follows the bones, not the object shape. I thought @RichardHaseltine, @Ominflux, @felis, and SickleYield might be interested to know what I found. And also they - or anyone else smarter than me - might disprove my theory if they know something I don't.  Which is a lot! 

Until then, I think I'll consider that one out of a thousand remaining mysteries solved. Thank you all!  

Matze

Hi John,
let me add my few cents about what I have learned so far.
I have experienced maybe the same "failor" that you have diskribed. See pic 1.
I guess I made a mistake when importing the geometrie into DAZ Studio. (in the Figure Setup Tab)
(I guess I used the wrong import values for the orientation of the mesh (-In your text above you wrote something similar).
When I rigged my figure a did my weight maps everything worked more or less as ist should.
But right from the start the Bounding Box was displaced..
When I tested my figure, trying to turn it arround its lenght (arround the y axis) -the figures root bone was selected, the figure actually turned arround the the wrong axis (x-axis I think) instead.
I took me a while to find out what went wrong... -your question here and the answers that you have got, have been a great help understanding the problem.
Finally when I selected the root bome in the joint editor, everythin made sense. For the Orientation of the Z axis there was a value of -90°!!!!
After I have corrected the Z orientation from -90° to 0°, the bounding box turned into the right position. See pic 2.
You can see the same effect when you translate the start or the end of the root bone. The bounding box will follow ant turn into the rigt position.

Maybe this explanation might be useful for someone triing his or her hands at rigging :-).

I would like to read more about rigging and weight maping. -the documentation for both processes could be more detailed :-)